US researchers develop tool to analyse baby cries to identify health problems

14 July 2013 (Last Updated July 14th, 2013 18:30)

A team of researchers from Brown University and Women & Infants Hospital of Rhode Island, US, has developed a computer-based tool that analyses the cries of infants, which will help doctors detect any neurological or developmental disorders at an early age.

A team of researchers from Brown University and Women and Infants Hospital of Rhode Island, US, has developed a computer-based tool that analyses the cries of infants, which will help doctors detect any neurological or developmental disorders at an early age.

A baby's cry is usually a sign of hunger, pain, or discomfort, but the team believes that the slight variations in the cry, which is not mostly perceptible to the human ears, could serve as a window into the brain and help in early intervention.

Brown University assistant professor of psychiatry and human behavior Stephen Sheinkopf said that there are lots of conditions that might manifest in differences in cry acoustics.

"Babies with birth trauma or brain injury as a result of complications in pregnancy or birth or babies who are extremely premature can have ongoing medical effects," Sheinkopf said.

"Cry analysis can be a noninvasive way to get a measurement of these disruptions in the neurobiological and neurobehavioral systems in very young babies."

The cry analysis tool, which operates in two phases, separates recorded cries into 12.5 millisecond frames during the first phase, where each frame is analysed on the basis of various parameters, including voicing, acoustic volume and frequency characteristics.

"Cry analysis can be a noninvasive way to get a measurement of these disruptions in the neurobiological and neurobehavioral systems in very young babies."

During the second phase, the tool considers the data taken from the first phase to provide a broader picture of the cry while also limiting the number of parameters to only those that are considered useful.

Then, the frames are put back together and considered either as an utterance or silence, with longer noiss then segregated from shorter ones and the time between the utterances.

Pitch and other variables are then averaged across each utterance and finally, the tool evaluates 80 different parameters, each of which could offer a view into an infant's health.

In order to identify what important aspects to look for, the university's laboratory for engineering man/machine systems director Harvey Silverman and his graduate students Brian Reggiannini and Xiaoxue Li worked closely with Sheinkopf and Brown's Center for the Study of Children at Risk director Barry Lester.

According to the researchers, neurological problems may alter the way infants control their vocal chords, and these small differences might appear in pitch and other acoustic features.

"Cry is an early warning sign that can be used in the context of looking at the whole baby," Lester said.

Sheinkopf intends to use the new tool to observe cry features that may correlate with autism.

The research was supported by the National Institutes of Health and the National Institute on Deafness and Other Communication Disorders.

The new tool was developed as part of a two-year collaboration between the faculties of Brown's School of Engineering and Women and Infants Hospital.


Image: Slight variations in a cry might indicate what's wrong with the baby's health. Photo: courtesy of Mike Cohea/Brown University.